International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1813

ISSN 2229-5518

Comparing methods of predicting pore pressure

Nwonodi Roland Ifeanyi

Abstract--- Comparative analysis uses certain criteria to judge various alternatives in the analysis. These criteria include the values of the alternatives
and the likelihood of obtaining them. The petroleum industry constantly seeks for ways to cut down cost of exploration and development. An improve-
ment on the ability to predict events ahead of the bit is the key to achieving this goal. Estimation of pore pressure finds itself so vital in the industry, es-
pecially in planning and drilling a modern deep well. No wonder there abound several methods for its prediction. The most common methods of predict-
ing pore pressure in the industry were compared in this study. Seismic travel time data for a well in the Louisianan basin was used for the comparison.
The methods compared were the equivalent depth, ratio, Eaton, Pennebarker, Hottman and Johnson, and, Matthew and Kelly. Based on the analysis in
the study, the method by Eaton is the most accurate.

Index Terms pore pressure, Seismic transit time, cost of exploration and development, Louisianan basin, comparative analysis, normal trend line,

effective stress.

1.0 INTRODUCTION

IJSER
Comparative analysis uses certain criteria to judge the various Predictive methods also called pre-drill

alternatives in the analysis. These criteria include the values of methods.

the alternatives and the likelihood of obtaining the values. Real time evaluation methods applied during

There are several methods for predicting pore pressure, yet drilling.

none of them are as accurate as making down-hole pressure Verification methods

measurement in the wellbore. Thus, the true expected value of In the pre-drill methods, predictions are based mainly on a

pore pressure in the formation is the value obtained by mak- combination of remote data such as seismic and basin model-

ing down-hole pressure measurement. Therefore, accurate ling, and/or by analysis of nearby wells [2]. In the case of

comparison of the alternatives is with respect to this value. planning a development well, emphasis is on data from previ-

Pore pressure is the pressure of fluids in the pores of a rock ous drilling experience in the area. Prediction from surface

formation. Pore pressure prediction is required in order to seismic data is by estimating seismic velocities, and then uti-

carry out safe drilling and well completion job. It is an integral lizing velocity-to-effective stress transform appropriate for the

process in the well planning procedure as well as the geologic given area. The estimated overburden stress combines with

evaluation of a potential trap. the effective stress to obtain pore pressure. The seismic data

Generally, techniques for predicting and detecting abnormal are also conditioned by gathering the signals and processing

formation pressure are often classified as follows [1]: them before they are used in the transform.

IJSER 2015
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1814
ISSN 2229-5518
In the real time evaluation methods, engineers monitor drill- portion of most of the sedimentary basins of the

ing and logging parameters while drilling the well. As drilling world, we can find abnormal pressure.

progresses into a transition zone, variations in rock properties

Table 1: Normal Pressure Gradient for geologic areas
and bit performance often provides many indirect indications
Source: Applied Drilling Engineering [1].
of changes in formation pressure [1]. Techniques used to esti-
Geologic areas Pressure gradients psi/ft
mate pore pressure real time includes:
Gulf coast 0.465
The d c exponent (which can also be used in predicting
Niger delta 0.433
pore pressure)
California 0.439
Measurement While Drilling and Logging While
Rocky mountains 0.436
Drilling
West Texas 0.442
Kicks

IJSER
North sea 0.452
other drilling rate factors

1.1 Types of pore pressure

1.2 Objectives of study
Practically, pore-pressure has been classed into the following
The aim of this study is to quantitatively scrutinize different
three groups:
methods of predicting pore pressure so as to note the similari-
1. Normal pressure: which is the hydrostatic pressure
ties and differences in them. It aims to determine the most
the fluids in the formation exert above a depth of in-
accurate of all the alternatives with respect to an expected val-
terest. It is usually expressed in terms of the hydro-
ue.
static gradient of the fluids in-situ. The normal pres-
1.3 Significance of study
sure gradients for several geologic areas having con-
The importance of this study cannot be overemphasized con-
siderable drilling activities are given in Table 1.
sidering the huge sums of money involved in hydrocarbon
2. Subnormal Pressure: This is an abnormally low pres-
exploration. For good drilling job, kicks and blowouts should
sure in the region.
be controlled at the same time that fracturing is monitored. A

3. Abnormal pressure: this is hydrostatic pressure high- blowout is characteristic of oil spill which degrades the envi-

er than the normal pressure in the zone. In at least a ronment. For example, in an offshore environment it can be a

threat to aquatic life which is a source of livelihood. Overpres-

IJSER 2015
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1815
ISSN 2229-5518
sure prediction is requisite for the design of drilling program, What are the key challenges in pore pressure predic-

casing design as well as choosing rig capacity. Pore pressure tions?

prediction helps to reduce well construction risk, save drilling

1.5 Scope of study

hour as well as cut down drilling cost. Thus, this study can be
The study includes methods for predicting pore pressure
applied in the following areas:
due to undercompaction phenomenon. It does not consider
Drilling engineering especially in planning and drill-
real time and post drill methods.
ing a modern deep well
2.0 LITERATURE
Well completion and work over operations in casing
Only a few comparative analyses on pore pressure prediction
designs and cement placement jobs
methods exist in literature. Although the industry has made
In control of wellbore for kicks and blowouts
undoubted progress in the understanding of pore pressure in
Reservoir engineering for material balance computa-

IJSER
recent years there are no fundamental change in the tech-
tions
niques used to predict pore pressure. The level of sophistica-
1.4 Statement of the problem
tion has indeed improved but there are still variations in pore
Accurate knowledge of pore pressure is a key requirement for

pressure predictions methods. To predict pore pressure, cer-

safe well planning in any kind of formation. There are several

tain techniques and correlations are employed.

methods for predicting pore pressure, and all of them are
2.1 Relationships used for pore pressure prediction
suppose to produce the same value of pore pressure in the
Most methods of predicting pore pressure are based on Ter-
rock. This is not usually the case as there are variations in the
zaghi`s effective stress principle [3], which implies that elastic
values obtained. Thus, this study wish to use real data to
wave velocities are a function of the effective stress tensor. The
compute pore pressure values using different methods so as
effective stress tensor is the difference between the total stress
to observe similarities and differences in them and also strife
tensor and the pore pressure. From Terzaghi`s work, the total
to know why this is so. Therefore, in carrying out this study,
stress equation is written as (1):
some basic questions are envisaged:
S = + P . . . . .. (1)
Have the existing methods lived up to performance in
S = the total stress, = the effective stress and P = the pore
predicting pore pressure?
pressure.
How best can pore pressure prediction methods be
From density-sonic log transform couple with estimate of av-
compared?
IJSER 2015
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1816
ISSN 2229-5518
erage sediment density from the top of the logged interval to 3
v
p = o ( o p n ) i . . .. . (2)
the seabed we derive the total stress [2]. vn

In addition, the effective stress is the principal driving mecha- Where P= predicted pore pressure, O= overburden pres-

nism for the compaction of compressible sediments. Under

sure, Pn = hydrostatic pressure, Vn = normally compacted
normal conditions, the magnitude of the effective stress in-
velocity and Vi = observed interval velocity from seismic data.
crease with depth and this invariably lead to a reduction in
2.2.2 Equivalent depth method
porosity. Laboratory studies [4], [5], confirm that effective

stress controls the compaction that take place in porous granu- This method also uses a reference normal compaction curve.

lar media. The procedure here is to compare the observed attribute with

2.2 Porosity-based methods the depth at which the observed attribute would at the normal

These methods rely on the normal compaction trend of sedi-

compaction curve [2]. This equivalent depth can then be used

IJSER
ments in a basin as far as it is derivable. Working with this
to compute the magnitude of the effective stress if the magni-

trend, it is assumed that compaction solely control porosity

tudes of the three principle stresses are known. The pore pres-

[2]. The formation is young, fine-grained sediments with simi-

sure is then calculated by subtracting the vertical effective

lar lithology and of low temperature. According to him, poros-

stress from the vertical stress at the depth of the observed at-

ity based pore pressure prediction method does not always

tribute. The implicit assumption is that porosity is controlled

deliver satisfactory results. This he said was either because

only by effective stress, driven by mechanical compaction. By

these assumptions are not valid, or because there are insuffi-

this method, the effective stress at the observed depth and that

cient data.
at the equivalent depth are the same. The formula can be given
2.2.1 Eaton`s method
as in (3):
The Eaton`s method compares an in-situ physical property to

a normally compacted equivalent physical property at the p = p n + ( o on ) . . . . . (3)

same depth [2]. It is valid as long as the normal curve can be

Where P = pore pressure in psi/ft, Pn = normal pore pressure
constructed for all depth of interest. This method typically

applies to seismic or acoustic velocity data and to resistivity at equivalent depth in psi/ft, O = overburden stress at depth

data. The form of the equation for velocity data is given below: in psi/ft, On = normal overburden at equivalent depth in

psi/ft.
IJSER 2015
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1817
ISSN 2229-5518
2.2.3 Ratio method

This method uses data on the seismic plot to determine pore

pressure. The model for predicting pore pressure can be given

by:

vn
Pp = Pn . . . . (4)
vi

2.3 Correlations used to estimate Pore pressure

Once the pressure gradient of a formation has been known, it

becomes very easy to predict the pressure existing at depth of

Fig.1. pressure gradient correlation by Pennebaker
interest. This is a quick guild to the estimation of pore pres-

IJSER
sure real time. Several researchers have used the interval tran- 2. Hottman and Johnson [7], gave a correlation between

sient time data to estimate the pressure gradient of formations interval transit time difference (t-t n ) and pore pres-

form where pore pressure is estimated. They include: sure using a Cartesian plot.

1. Pennebaker [6], used the transient time ratio of (t/t n )

to estimate pressure gradient. From this, pore pres-

sure can be estimated.

Fig.2. pressure gradient correlation by Hottman and Johnson

IJSER 2015
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1818
ISSN 2229-5518
3. Mathews and Kelly [8], gave a relation between the ods. They ranked Eaton method as the most accurate.

difference in interval transit time, (t-t n ) and pressure

gradient but used a semi-log plot. 3.0 METHODOLOGY

The petroleum industry constantly seeks for ways to cut down

cost. An improvement on the ability to predict events ahead of

the bit is a key to achieving this goal. Prediction of pore pres-

sure finds itself so vital in the industry especially in planning

and drilling a modern deep well, no wonder there abound

several methods used to get it.

This section focuses on the method adopted in this study.

IJSER
Fig.3. pressure gradient correlation by Mathews and Kelly

Eaton et al [9], investigated currently applied technologies

3.1 Data collection
used to predict, detect, and evaluate the magnitudes of ab-

normal pressure and fracture gradient in the earth crust. They The source of data for analysis in this study is through sec-

carried out a wide literature search which covered most of the ondary data sourcing. The use of seismic data from a well off-

published articles concerning the subject. They made a de- shore Louisiana aided the analysis.

tailed interview of Thirty (30) companies which included in-

3.2 Method of analysis

ternational major oil companies, drilling contractors and ser-

From the interval transit time, pore pressure at a depth of
vice companies from the US, Canada and South America. They
19,000 ft was predicted using different methods. After this, the
also used actual well and formation data to evaluate pore
arithmetic average of the various alternatives was calculated.
pressure and fracture gradient at various depths. They went
This value was taken as the expected value of the pore pres-
further and statistically determined the methods from the in-
sure in the formation. Afterwards, statistical analysis of the
terview. According to them, the best methods employed by
results was carried out by calculating the absolute deviation
drilling personnel for pore pressure prediction are the
from the mean.
Hottman and Johnson, Equivalent depth and the Eaton meth-

IJSER 2015
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1819
ISSN 2229-5518
3.3 Accuracy of the models

The fraction of absolute error, e, was used as the measure of

merit. It was calculated by (5):

xx
e=
. . . . . (5)
x

x = the value used and x = the mean of all the values used.

The original pore pressure from down-hole measurement was

not given as a priori and so the calculated mean was used for

IJSER
the analysis.

The alternative that gave value closest to the expected value The normal pressure gradient in the basin is 0.465 psi/ft.

was ranked as most accurate. This alternative is the one hav- Equivalent depth is at 2500 ft and overburden gradient is 0.995

ing the least fraction of absolute error. psi/ft.

4.0 ANALYSIS OF RESULTS

Table 2 shows the well data obtained from the Louisianan ba-

sin.

Table 2. Well data from Louisiana basin [1]. (Adapted from

tamu-pemex, pore pressure prediction.ppt, slide 18)

Fig.4. depth vs transit time plot

IJSER 2015
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1820
ISSN 2229-5518
4.3 Analysis of result The method by Hottman and Johnson, and that by Mathews

and Kelly both use the difference between the interval transit
Computation of results is given in the excel spreadsheet that

follows:
time at the needed point and that at the normal line, to tn ,

to compute the pressure gradient. Yet the values obtained

COMPUTATIONS FOR GULF COAST WELL
Data Input from them are different. This difference is probably due to the
o 0.995 psi/ft To 95 sec/ft Z 19000 ft
assumptions used in generating the correlations. This can be a
Pn 0.465 psi/ft Tn 60 sec/ft Ze 2500 ft
major challenge in pore pressure prediction methods. Some
Equivalent Depth Method Ratio Method Eaton`s Method
Pne 1162.5 psi to 95 sec/ft to 95 sec/ftare valid while others are definitely not.
oe 2487.5 psi tn 60 sec/ft tn 60 sec/ft
o 18905 psi pn 8835 psi
The ratio method gave a very different result, which underes-
p 17580 psi p 13988.75 psi
p 0.925263 psi/ft p 0.73625 psi/ft p 0.861476 psi/ft
timates the mean value. Compared with the Eaton`s method,

IJSER
Pennebarker correlation Hottman &Johnson Matthew & Kelly there seems to be a significant error in the development of the
to/tn 1.583333 to-tn 35 sec/ft to-tn 35
p=f(to/tn) 0.97 psi/ft p=f(to-tn) 0.9 psi/ft p=f(to-tn) 0.8 psi/ft model.
Mean value 0.865498 Psi/ft
Selection of a slightly different normal compaction trend

shows that the equivalent depth changes to 2000ft with a nor-

mal transit time of 65 sec/ ft . With this, the various values

4.4. Discussion of results
obtained are given next.

The results show that the values are not all the same. For ex-

ample, the ratio method gives a value of 0.68 psi/ft while the

Hottman and Johnson method gives a value 0.87 psi/ft. The

difference between these results is quite significant and can be

costly to drilling job. It is indicative of an error somewhere.

By use of the values obtained from the various alternatives, the

most accurate method is that obtained by use of the Eaton`s

method. It is closest to the average value.

IJSER 2015
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1821
ISSN 2229-5518
COMPUTATIONS FOR GULF COAST WELL Based on the results obtained in this study, the following can
Data Input
be drawn
o 0.995 psi/ft To 95 sec/ft Z 19000 ft
Pn 0.465 psi/ft Tn 65 sec/ft Ze 2000 ft
Porosity based pore pressure prediction models does
Equivalent Depth Method Ratio Method Eaton`s Method
not give accurate results unless an appropriate nor-
Pne 930 psi to 95 sec/ft to 95 sec/ft
oe 1990 psi tn 65 sec/ft tn 65 sec/ft mal compaction curve has been defined.
o 18905 psi pn 8835 psi
p 17845 psi p 12912.69 psi
The inability to accurately model deviations from
p 0.939211 psi/ft p 0.679615 psi/ft p 0.825236 psi/ft
normal trend line is the major reason while different
Pennebarker correlation Hottman &Johnson Matthew & Kelly
methods of pore pressure prediction do not give
to/tn 1.461538 to-tn 30 sec/ft to-tn 30
p=f(to/tn) 0.94 psi/ft p=f(to-tn) 0.87 psi/ft p=f(to-tn) 0.77 psi/ft equal results.
Mean value 0.837344 Psi/ft

IJSER
5.2 Recommendations
This reveals that only the equivalent depth method has its val-
Porosity based models should be employed only
ue increased while other methods have their values reduced.
when appropriate normal compaction curves are de-
Similar shift in the normal compaction trend in the opposite
fined for the field
direction reveals an increase in values for other methods, but a

decrease in value for the equivalent depth method. The equiv- Offset wells should always be used to calibrate pre-

alent depth method tend to overestimate or underestimate diction results

pore pressure based on whether the equivalent depth is to-

Direct pressure measurements should be taken in the
ward the surface or towards the target.
life of the well so as to check for the consistencies of

All the methods have a common value, which equals the nor- models used to get pore pressure results.

mal pressure gradient, at the point where the equivalent depth

For seismic data, the use of the equivalent depth
equals the depth of interest and the normal transit time equals
method should be a last resort.
the recorded transit time. Outside this point, there seems to be

an inability to accurately model deviations from normal trend. ACKNOWLEDGEMENT

5.0 CONCLUSIONS AND RECOMMENDATION

Many thanks go to the authors who have already done com-

5.1 Conclusions parative study on pore pressure prediction methods. Great

IJSER 2015
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 7, July-2015 1822
ISSN 2229-5518
insight was obtained in their work. [7] C.E., Hottman, R.K., Johnson, Estimation of formation

pressures from log derived shale properties, JPT, vol 6, June pp 717-
ABOUT THE AUTHOR
722, 1965.

The author is currently pursuing a masters degree in P etrole-

[8] W.R., Mathews, J., Kelly, How to Predict Formation
um Engineering in the University of Port Harcourt. Email ad-
Pressures and Fracture Gradient from Electric and Sonic Logs,
dress: rolandnwonodi@yahoo.com
Oil and Gas journal, Feb, 20. 1967.

REFERENCES
[9] B.A., Eaton, S., Ikeda, C.,Yolshida,An investigative

[1] A.T., Bourgoyne, K.K., Jr., Millheim, M.E., Chenevert, Study of Recent Technology Use for Prediction, Detection, and

F.S., Young, Applied Drilling Engineering, Society Of Petrole- Evaluation of Abnormal formation pressure and Fracture Gra-

um Engineers, Textbook Series, v.5, pp 246-299, 1991. dient in North and South America, IADC/SPE 36381, pp 131-

IJSER
151, 1996.
[2] R.E., Swarbrick, Challenges of porosity-based pore

pressure prediction, CSEG Recorder, Sep, p 74, 2002.

[3] K., Terzaghi, Theoretical soil mechanics, John Wiley

and Sons, Inc, 1947.

[4] C.A., Tosaya, Acoustical properties of clay-bearing

rocks, PhD Dissertation, Stanford University Department of

Geophysics, 1982.

[5] J., Dvorkin, D., Moos, J.L., Packwood, and A.M., Nur,

Identifying patchy saturation from well logs, geophysics, 64.

pp 1756-1759, 1999.

[6] E.S, Pennebaker, Seismic data indicate depth and

magnitude of abnormal pressure, world oil, vol166, pp 73-82,

1968.

IJSER 2015
http://www.ijser.org